MI-Net: A Deep Network for Non-linear Ultrasound Computed Tomography Reconstruction

TitleMI-Net: A Deep Network for Non-linear Ultrasound Computed Tomography Reconstruction
Publication TypeConference Paper
Year of Publication2020
AuthorsFan, Y, Wang, H, Gemmeke, H, Hesser, J
Conference Name2020 IEEE International Ultrasonics Symposium (IUS)
Date Publishedsep
Keywordsacoustic tomography, Acoustics, biomedical ultrasonics, breast cancer detection, cancer, classical iterative optimization strategies, Computed tomography, convolutional neural nets, convolutional neural network, deep network, frequency pressure field, high-resolution imaging method, image reconstruction, image resolution, inverse problem, inverse problems, iterative methods, iterative solution, medical image processing, MI-net, non-linear forward model, nonlinear ultrasound computed tomography reconstruction, OA-Breast, Optical and Acoustic Breast Phantom Database, Phantoms, reconstruction time, Three-dimensional displays, time consumption problem, Time-frequency analysis, Training, ultrasonic imaging, USCT, USCT data, USCT image reconstruction, wave equation
Abstract

Ultrasound Computed Tomography (USCT) is a new high resolution imaging method with high potential for breast cancer detection. USCT image reconstruction requires the iterative solution of a wave equation and its adjoint, being very time-consuming. To address this time consumption problem, a convolutional neural network with frequency pressure field as input is presented, outperforming classical iterative optimization strategies to solve the inverse problem. The Optical and Acoustic Breast Phantom Database (OA-Breast) is used to generate simulated USCT data using paraxial approximation forward model both for training and testing. Extensive experiments demonstrate that our neural network outperforms the classical strategies both in reconstruction time and quality.

DOI10.1109/IUS46767.2020.9251441
Citation Keyfan_mi-net_2020